|Home | About | Journals | Submit | Contact Us | Français|
With aging, multiple Ca2+-associated electrophysiological processes exhibit increased magnitude in hippocampal pyramidal neurons, including the Ca2+-dependent slow afterhyperpolarization (sAHP), L-type voltage-gated Ca2+channel (L-VGCC) activity, Ca2+-induced Ca2+ release (CICR) from ryanodine receptors (RyRs), and Ca2+ transients. This pattern of Ca2+ dysregulation correlates with reduced neuronal excitability/plasticity and impaired learning/memory and has been proposed to contribute to unhealthy brain aging and Alzheimer’s disease. However, little is known about the underlying molecular mechanisms. In cardiomyocytes, FK506-binding protein 1b/12.6 (FKBP1b) binds and stabilizes RyR2 in the closed state, inhibiting RyR-mediated Ca2+release. Moreover, we recently found that hippocampal Fkbp1b expression is downregulated, whereas Ryr2 and Frap1/Mtor (mammalian target of rapamycin) expression is up-regulated with aging in rats. Here, we tested the hypothesis that disrupting FKBP1b function also destabilizes Ca2+ homeostasis in hippocampal neurons and is sufficient to induce the aging phenotype of Ca2+ dysregulation in young animals. Selective knockdown of Fkbp1b with interfering RNA in vitro (96 h) enhanced voltage-gated Ca2+ current in cultured neurons, whereas in vivo Fkbp1b knock-down by microinjection of viral vector (3–4 weeks) dramatically increased the sAHP in hippocampal slice neurons from young-adult rats. Rapamycin, which displaces FKBP1b from RyRs in myocytes, similarly enhanced VGCC current and the sAHP and also increased CICR. Moreover, FKBP1b knockdown in vivo was associated with upregulation of RyR2 and mTOR protein expression. Thus, disruption of FKBP1b recapitulated much of the Ca2+-dysregulation aging phenotype in young rat hippocampus, supporting a novel hypothesis that declining FKBP function plays a major role in unhealthy brain aging.
Substantial evidence indicates that, with aging, multiple Ca2+-associated electrophysiological processes are increased in magnitude in hippocampal pyramidal neurons, including the Ca2+-dependent slow afterhyperpolarization (sAHP), the Ca2+ action potential (AP), L-type voltage-gated Ca2+ channel (L-VGCC) activity, Na+ action potential failure (accommodation), synaptic depression, Ca2+-induced Ca2+ release (CICR) by the Ca2+-sensing/releasing ryanodine receptors (RyRs) of the smooth endoplasmic reticulum, and voltage-activated Ca2+ transients (Landfield and Pitler, 1984; Moyer et al., 1992; Disterhoft et al., 1996; Thibault and Landfield, 1996; Norris et al., 1998; Thibault et al., 2001, 2007; Clodfelter et al., 2002; Hemond and Jaffe, 2005; Kumar and Foster, 2005; Gant et al., 2006). This pattern of aging-related Ca2+ dysregulation has been correlated with impaired plasticity and learning/memory (Disterhoft et al., 1996; Thibault and Landfield, 1996; Tombaugh et al., 2005; Tonkikh et al., 2006; Disterhoft and Oh, 2007). Furthermore, increased neuronal RyR activity also appears to develop in transgenic mouse models of Alzheimer’s disease (AD) (Stutzmann et al., 2007; Bezprozvanny and Mattson, 2008). Conversely, hippocampal long-term potentiation (LTP), which depends more on Ca2+ influx via NMDA receptors than VGCCs, is generally decreased or unchanged in aged animals (Watabe and O’Dell, 2003; Kumar and Foster, 2004; Burke and Barnes, 2006; Lynch et al., 2006; Magnusson et al., 2006; Smith et al., 2009). Altered Ca2+ regulation with aging, although of different patterns, is also found in nonpyramidal neurons (Michaelis et al., 1984; Gibson and Peterson, 1987; Reynolds and Carlen, 1989; Kirischuk et al., 1996; Murchison and Griffith, 2007). Together, these and other findings in aged animals have suggested the hypothesis that Ca2+ dysregulation is a major factor in unhealthy brain aging and AD(Gibson and Peterson, 1987; Landfield, 1987; Khachaturian, 1989; Disterhoft et al., 1996; Toescu and Verkhratsky, 2003; Disterhoft and Oh, 2007; Murchison and Griffith, 2007; Thibault et al., 2007).
Nevertheless, little is known about the molecular mechanisms that underlie aging changes in Ca2+ regulation in pyramidal neurons. In myocytes, the immunophilins FK506-binding protein 1a and 1b (FKBP1a and FKBP1b, also called FKBP12 and FKBP12.6) normally bind and stabilize RyR channel complexes in the closed state, thereby inhibiting intracellular Ca2+ release (Lehnart et al., 2003; Zalk et al., 2007). In particular, FKBP1b binds with high affinity to RyR2, the most abundant RyR isoform in brain and heart. In recent microarray studies, we found that expression levels of Fkbp1a and Fkbp1b are downregulated with aging in hippocampus of rats (Kadish et al., 2009, their Table 3). Thus, we tested here whether disrupting FKBP1b function in hippocampal neurons of young animals might be sufficient to recapitulate substantial aspects of the aging phenotype of Ca2+ dysregulation (Gant et al., 2006).
To disrupt FKBP1b function, we used in vitro and in vivo selective knockdown of Fkbp1b expression with small interfering RNA (siRNA) and treatment with rapamycin, an immunosuppressant drug that binds and displaces FKBP1b from RyR2 in myocytes, increasing intracellular Ca2+ release (Marks, 1997; Long et al., 2007; Zalk et al., 2007). Our results support a major role for FKBPs in the hippocampal aging phenotype.
Studies on adult animals used 3- to 4-month-old male Fischer 344 rats obtained from Harlan. All protocols and procedures were performed in accordance with institutional guidelines and were approved by the Animal Care and Use Committee.
Mixed neuronal–astrocyte cultures were prepared as described previously (Porter et al., 1997; Clodfelter et al., 2002; Norris et al., 2002) and transfected with 8 nm small interfering RNA against Fkbp1bmRNA(siFkbp1b) (58950, FKBP1b siGENOME SMART-pool; Dharmacon), 8 nm nontargeted siRNA (ON-TARGETplus Non-targeting siRNA #1; Dharmacon), or siRNA dilution buffer on 7 or 8 d in vitro (DIV). Using the SilenceMag protocol of MagnetoFection technology (OZ Biosciences), 8 nm of either siFkbp1b or nontargeting siRNA was mixed with 4 µl of SilenceMag beads in 200 µl of serum-free supplemented MEM at room temperature for 20 min and then added to each well of six-well hippocampal cell culture plates to a final volume of 2 ml. The cell cultures were then placed on a magnetic plate designed for magnetic transfection (OZ Biosciences) for 15 min at room temperature, after which they were placed in a 37°C cell culture incubator until analyses. Cultures were treated with siFkbp1b, nontargeting siRNA, or siRNA buffer for a period of ~4 d, followed by recording and quantitative PCR (qPCR) analysis, performed on cells plated on “micro-islands” (Blalock et al., 1999).
To achieve selective knockdown in vivo, an adeno-associated virus (AAV) vector was used to express a short hairpin RNA (shRNA) targeting the rat Fkbp1b mRNA (GenBank accession no. NM_022675). The construction of this vector (shFkbp1b) was performed at the University of Pennsylvania Vector Core by first cloning a double-stranded fragment generated by annealing two oligonucleotides (top strand, 5′-CGGAAGGACATTCCCTAAG_TTCAAGAGA_CTTAGGGAATGTCCTTCCGC-3′; bottom strand, 3′-GCCTTCCTGTAAGGGATTC_AAGTTCTCT_GAATCCCTTACAGGAAGGCG-5′) into the AAV pZacf–U6–ZsGreen plasmid (provided by the University of Pennsylvania Vector Core) at the BamH1–EcoR1 restriction enzyme sites. The recombinant clone was then used to produce the AAV1 serotype shFkbp1b vector (AAV1–shFkbp1b). This design places the shRNA under the control of the human U6 promoter, which is followed by a downstream cytomegalovirus promoter-driven ZsGreen fluorescence marker. When cleaved, the shRNA was designed to generate siRNA targeting the CGGAAGGACATTCCCTAAG region of rat Fkbp1b mRNA. Purification and determination of AAV1–shFkbp1b vector titers were performed by the University of Pennsylvania Vector Core.
In recent years, considerable progress has been made in developing the techniques for viral-vector-mediated gene transfer (Fitzsimons et al., 2002; Klein et al., 2002, 2008; Zaiss and Muruve, 2005; Cearley and Wolfe, 2006; Mouravlev et al., 2006). Choice of the AAV1 serotype for specific transduction of targeted pyramidal neurons in the hippocampus in the present study was based on the previous work and on a series of pilot studies in which we compared distribution of neurons expressing green fluorescent protein (GFP) after injection of GFP-expressing AAV serotypes 1, 8, and 9. Under isoflurane anesthesia, rats were placed in a David Kopf Instruments stereotaxic frame. Small holes were drilled in the skull either unilaterally or bilaterally (at points from bregma: posterior, −4.5 mm; lateral, 3.0 mm), and the dura was pierced to allow for microinjection. A Hamilton microsyringe was lowered 1.6 mm into the dorsal hippocampus, a region critical for spatial memory (Moser and Moser, 1998). At the target injection site (CA1 stratum oriens at the peak of the curve of stratum pyramidale), 2 µl of AAV-containing vehicle was delivered at a rate of 0.2 µl/min using a Stoelting QSI microinjection pump. After injection, the syringe was left in place for 5 min, then the small holes in the skull were filled with bone wax, and the incision was sutured. Two to 4 weeks after unilateral microinjections of AAV, rats were killed, and the brains prepared for sectioning and microscopy. GFP expression showed that all serotypes successfully transduced hippocampal CA1–CA3 pyramidal neurons and that expression of GFP was primarily limited to the hippocampus proper. Serotypes 1 and 9 but not 8 exhibited longitudinal spread sufficient to transduce neurons throughout the dorsal regions of the ipsilateral hippocampus and did not spread substantially across either the hippocampal fissure into the dentate gyrus or the midline into the contralateral hippocampus. Serotype 1 appeared slightly more specific for the somal layer and was selected as the preferred serotype for additional studies. For this serotype, we estimated from hematoxylin–eosin-counterstained sections that 30–60% of CA1–CA3 pyramidal neurons in ipsilateral dorsal hippocampus expressed GFP.
For the electrophysiological experiments, one hippocampus (ipsilateral) of 3-month-old male F344 rats was microinjected with 2 µl of AAV [titer, 2.2E13GC(genome copies per milliliter)] expressing short hairpin RNA against Fkbp1b (shFkbp1b), whereas the control hippocampus of each animal (contralateral) was either microinjected with 2 µl of empty-vector control AAV (AAV titer, 3.56E12 GC) or received no treatment. Three to 4 weeks after injection, intracellular recordings of the sAHP were obtained from CA1 pyramidal neurons in acute hippocampal slices from both ipsilateral and contralateral sides of nine shFkbp1b-injected animals. Subsequently, qPCR of Fkbp1b was performed on tissue from the dorsal tips and one slice each from both hippocampi of eight of the nine animals studied electrophysiologically. In addition, brains of two other animals unilaterally injected in hippocampus with shFkbp1b were prepared for immunohistochemical studies (see below).
For mRNA quantification, one-step real-time reverse transcription (RT)-PCR was used. RT-PCR amplification was performed as described previously (Chen et al., 2000) using an ABI prism 7700 sequence detection system (Applied Biosystems) and TaqMan One Step RT-PCR reagents (Applied Biosystems). All samples were run in duplicate in a final volume of 50 µl containing 50–100 ng of cellular RNA and a Taq-Man probe (200 µm) and primers (300 nm each) with an amplicon spanning the rat Fkbp1b cDNA region from nucleotides 155 to 259. Cycling parameters for all assays were as follows: 30 min at 48°C, 10 min at 95°C followed by 40 cycles of 15 s at 95°C, and 1 min at 60°C. Primers were designed using Primer Express software (version 1.5; Applied Biosystems) and chemically synthesized by Applied Biosystems (forward primer, 5′-GCAAGCAGGAAGTCATCAAAGG-3′; reverse primer, 5′-CAGTAGCTCCATATGCCACATCA-3′; TaqMan probe, 5′-AGCTCATCTGGGCAGCGCCTTCTT-3′). The RNA levels of glyceraldehyde-3-phosphate dehydrogenase were used as normalization controls for RNA quantification.
Immunohistochemistry (IHC) methods were similar to those described previously (Kadish and Van Groen, 2002). Rats were anesthetized and perfused transcardially with cold 0.9% NaCl, followed by perfusion with 4% paraformaldehyde in PBS; thereafter, they were stored overnight in the 4% paraformaldehyde/PBS solution and then transferred to 30% sucrose/PBS solution. Coronal sections (30µm) were cut on a freezing sliding microtome. The following primary antibodies were used for overnight incubation: FKBP12.6/1b (1:500; sc-98742; Santa Cruz Biotechnology), mammalian target of rapamycin (mTOR) (1:500; ab32028; Abcam), and RyR2 (1:150; AB9080; Millipore Corporation). Sections were then rinsed and transferred to a solution containing appropriate biotinylated secondary antibody for 2 h, after which they were rinsed and transferred to a solution containing Extra-Avidin for 2 h. Sections were then incubated for 3 min with Ni-enhanced DAB solution. To obtain comparable staining across sections, all animals were processed simultaneously in the same staining tray. The FKBP1b/12.6-, mTOR-, and RyR2-stained sections of the dorsal hippocampus were digitized using an Olympus DP70 camera, and the resulting images were analyzed using the ScionImage (NIH Image) program. The DP70 camera settings were maintained throughout acquisition of all photomicrographs for statistical comparison. For densitometric analysis of FKBP1b, mTOR, and RyR2 immunohistochemistry, the CA1 pyramidal cell soma layer (stratum pyramidale) of the ipsilateral and contralateral hippocampus on two sections from each of two animals were outlined and quantified.
Rapamycin (A.G. Scientific) was dissolved in DMSO (0.1% final DMSO) and added to the culture dish (whole-cell recording) or holding chamber (intracellular recording in slices) to a final concentration of 5 µm. Rapamycin exposure was maintained for 1, 24, or 96 h in cultures and for 2–6 h in slices before recording began. Nimodipine was dissolved in DMSO and added to cultures at a final concentration of 10 µm (0.1% final DMSO dilution). Ryanodine (Alomone Labs) was dissolved in water and perfused into the recording chamber at a rate 1000 times slower than the oxygenated artificial CSF (ACSF) perfusate. Final concentration of ryanodine in the recording chamber was estimated at 20µm.
Whole-cell recordings of pharmacologically isolated Ca2+ channel currents, using Ba2+ as the charge carrier, were obtained from neurons in primary hippocampal cell cultures at 9–12 DIV, as described previously (Porter et al., 1997; Blalock et al., 1999; Brewer et al., 2001; Norris et al., 2002). Electrophysiological data were acquired at 5–20 kHz using a patch-clamp amplifier (Axopatch 200A), DigiData 1320 digital input/output board and pClamp 7 (Molecular Devices). Cell capacitance measures, membrane resistance, access resistance, and holding current were obtained from the membrane properties function in pClamp7. All cells were held at −70 mV unless stated otherwise. Current–voltage (I–V) relationships were determined by stepping the voltage in increments of 10 mV from 70 to +60 mV for 150 ms. Current was averaged over five traces for each measurement point, and current density was determined by dividing current by cell membrane capacitance (picoamperes per picofarad).
Young-adult male F344 rats were anesthetized in a CO2 chamber before rapid decapitation. Our sharp-electrode electrophysiological methods have been described previously (Thibault et al., 2001; Gant et al., 2006). Briefly, intracellular recordings were obtained from CA1 pyramidal neurons in hippocampal slices (350-µm-thick) maintained in oxygenated ACSF (in mm: 128 NaCl, 1.25 KH2PO4, 10 glucose, 26 NaHCO3, 3 KCl, 2 CaCl2, and 2 MgCl2), using sharp glass pipettes filled with HEPES (10 mm), KMeSO4 (2 m), and Calcium Orange tetrapotassium salt (5 mm; Invitrogen). Electrophysiological data were acquired and analyzed using pClamp 8, a sharp-electrode amplifier (Axoclamp 2B), and a DigiData 1320 board (Molecular Devices). Voltage records were digitized at 2–20 kHz and low-pass-filtered at 1 kHz. Input resistance and AHP measures were obtained in current-clamp mode. AHP duration and sAHP amplitude were measured after four Na+ APs triggered by an intracellular depolarizing pulse. The 5 mm concentration of Ca2+ indicator used is in a range commonly used in slice imaging experiments (Jaffe et al., 1992; Brown and Jaffe, 1994; Jaffe and Brown, 1994) and apparently does not distort AHP waveforms or obscure aging differences in the AHP (Thibault et al., 2001; Gant et al., 2006).
Methods used were similar to those described previously by several groups (Miyakawa et al., 1992; Brown and Jaffe, 1994; Jaffe and Brown, 1994; Magee and Johnston, 1995; Thibault et al., 2001; Gant et al., 2006). Individual CA1 neurons impaled with sharp pipettes and loaded with the Ca2+ indicator Calcium Orange were imaged on the stage of a Nikon E600 microscope equipped with a 40× water-immersion objective and a CCD camera (Roper Scientific/Princeton Instruments). Calcium Orange was allowed to diffuse into the cell for at least 10 min before Ca2+ fluorescence measures were performed. Calcium Orange was excited using a wavelength switcher (Sutter Instruments Lambda DG-4) and software control (Axon Imaging Workbench; version 184.108.40.206; Molecular Devices). The 575 nm wavelength was monitored through a dichroic mirror centered at 570 nm during excitation with the 550 nm wavelength. The intracellular Ca2+ response was elicited by 10 s of 7 Hz repetitive synaptic stimulation (RSS) delivered to the Schaffer collaterals via bipolar stimulation electrode. Stimulation intensity was set at AP threshold and generated an action potential on essentially each pulse (Thibault et al., 2001). Ca2+ responses were measured in a region of interest corresponding to the visible outline of the cell soma, excluding surrounding low-intensity diffracted light. Although the central region of the neuron appears brighter (see Fig. 6, inset), this may be related in part to greater indicator loading in this thicker central region of the cell (Tsien, 1988; Nicotera et al., 1994). After background subtraction from an area devoid of cells adjacent to the recorded cell, percentage change in fluorescence was determined relative to baseline (%ΔF/F). Optical sampling rate averaged ~5 Hz, which appeared sufficiently sensitive to detect significant changes in the Ca2+ response. Excitation of Calcium Orange was essentially continuous during a 10 s train of RSS, but control runs showed that bleaching was minimal (−0.21 ± 0.09% at 5 s RSS;−0.69 ± 0.35% at 10 s RSS; n=7 cells). The ryanodine-sensitive component (CICR component) of the Ca2+ response was determined by subtracting the post-ryanodine Ca2+ response from the pre-ryanodine response [(pre-ryanodine%ΔF/F) − (post-ryanodine%ΔF/F) = ryanodine-sensitive component].
Data analysis was performed with Clampfit8 (Molecular Devices), and statistical analysis was performed with Stat-View (SAS Institute). Variables were analyzed using ANOVA across all groups with means and SEs reported. Fisher’s protected least significant difference (PLSD) test was used for post hoc group comparisons. p values <0.05 were considered significant.
Because increased L-VGCC activity is a primary component of aging-related Ca2+ dysregulation (Campbell et al., 1996; Thibault et al., 2001, 2007), we tested the effects of prolonged exposure to small interfering RNA against Fkbp1b (siFkbp1b) or to rapamycin on pharmacologically isolated VGCC currents recorded in whole-cell patch-clamp mode from neurons in primary hippocampal cell cultures. Whole-cell recording data were analyzed from a total of 149 neurons across nine treatment groups, including untreated (control) cells and cells treated with siFkbp1b (8 nm, 96 h), nontarget control siRNA (8 nm, 96 h), siRNA buffer vehicle (96 h), rapamycin 5µm (1, 24, and 96 h), rapamycin (5µm, 96 h) plus the selective blocker of L-type VGCCs, nimodipine (10µm), and nimodipine alone (10µm). Neurons in all the different treatment conditions were recorded alternately and treatment groups were run concurrently over the length of the study, to control for any drift in experimental conditions. Table 1 provides values for basic electrophysiological properties and parameters, which revealed no significant differences across treatments on any of these variables.
Figure 1 shows representative current traces, I–V curves, and mean peak current density for treatment groups in these experiments. An ANOVA across all groups revealed a main effect of treatment (F(8,140) = 18.49, p < 0.0001) on peak Ca2+ current density (Fig. 1C). Fisher’s PLSD post hoc test showed that, after 96 h incubation in siFkbp1b, VGCC current density was significantly enhanced versus controls (p < 0.0001). Cultured neurons treated 96 h with either nontarget siRNA or siRNA vehicle exhibited no effects on VGCC current density (Fig. 1C, right bars). Real-time qPCR showed that the siFkbp1b transfection treatment achieved significant knockdown of cell culture Fkbp1b mRNA (mean ± SEM for siFkbp1b-treated vs control, 0.91 ± 0.05 vs 1.15 ± 0.05, p < 0.01).
Rapamycin exposure for 1 h had no effect on VGCC current density compared with controls, consistent with a previous study showing little effect of acute rapamycin exposure (Norris et al., 2002). However, after 24 or 96 h incubation with rapamycin, neurons exhibited substantially larger VGCC currents (p < 0.001, 24 or 96 h incubation duration vs controls) (Fig. 1C, left bars). Despite significant effects on current density, neither siFkbp1b nor rapamycin altered current inactivation during the voltage step (Fig. 1A) or shifted the I–V relationship (Fig. 1B). To determine whether the effect of rapamycin was mediated by actions specific for L-VGCCs, we added the specific L-VGCC blocker nimodipine (10 µm) to some cultures. Notably, nimodipine fully blocked the enhancing effect of rapamycin on total VGCC current (p < 0.0001), reducing current density levels in rapamycin-treated cells to those seen in untreated controls (Fig. 1C, middle bars). Furthermore, in cells treated with nimodipine alone, VGCC current density was significantly reduced by ~30–35% compared with controls (p < 0.05), consistent with values previously reported for the L-type current fraction in hippocampal neurons (Blalock et al., 1999). As described previously, although slowly inactivating L-VGCCs are high voltage activated, nimodipine blockade shifts the peak of the I–V curve to more positive potentials (Fig. 1B) (Tsien et al., 1988; Regan et al., 1991; Porter et al., 1997; Blalock et al., 1999). Thus, most of the enhancing effect of rapamycin on VGCCs appeared to be accounted for by actions specific to L-type channels. Moreover, the similarities in the inactivation patterns and I–V curves of siFkbp1b-treated and rapamycin-treated neurons (Fig. 1A,B) indicate that these two treatments acted primarily on the same type (L-VGCC) of VGCC.
In these studies, we tested whether FKBP1b knockdown in adult neurons in vivo could increase sAHP magnitude because a larger sAHP is one of the most consistent electrophysiological markers of aging-related hippocampal Ca2+ dysregulation in pyramidal cells (Disterhoft and Oh, 2007; Thibault et al., 2007). AHP measures were obtained from nine young rats 3–4 weeks after unilateral hippocampal microinjection of shFkbp1b. The AHP was recorded intrasomatically from 17 CA1 neurons in slices from the hippocampus ipsilateral to the FKBP1b knockdown injection and from 17 CA1 neurons of the control (contralateral) hippocampus (Fig. 2, top). Fluorescence microscopy indicated that, 3–4 weeks after injection, the viral vector had infected neurons throughout much of CA1–C3 of the ipsilateral dorsal hippocampus. The AAV vector did not appear to spread across the midline to the contralateral hippocampus, although faint fluorescence was often seen in the contralateral CA1 stratum oriens and dentate gyrus (Fig. 2, middle), possibly reflecting transport via commissural fibers from the ipsilateral pyramidal neurons or uptake of AAV particles leaked into the ventricle. Nevertheless, neuronal somata were not generally infected in the contralateral hippocampus. Statistical analysis revealed that mean sAHP amplitude and duration approximately doubled in neurons from the FKBP1b knockdown side (Fig. 2). In addition, 9 of 17 knockdown side neurons showed sAHP amplitude or duration values above the highest control value. The sAHP magnitudes of control neurons were similar to those observed in young rats, whereas values from knockdown side neurons were similar to those observed in aged rats in previous studies (Gant et al., 2006).
qPCR (Fig. 3) and IHC densitometry (Fig. 4, top; Table 2) results from this knockdown study clearly showed that FKBP1b knockdown was achieved at both the mRNA (25–35%) and protein (Table 2) levels in the ipsilateral hippocampus. Moreover, the results revealed apparent upregulation of protein levels for two targets of FKBP inhibition, RyR2 and mTOR (see Discussion), on the FKBP1b knockdown side (Fig. 4; Table 2). Therefore, the present data suggest that FKBP1b may repress RyR2 and mTOR expression through an as yet unknown signaling pathway.
In other experiments on acute hippocampal slices of young-adult F344 rats, we tested whether rapamycin also could induce aging-like effects on the sAHP. For these studies, we again used intracellular sharp electrode recording techniques in individual CA1 pyramidal neurons. Slices were incubated in either control medium or rapamycin (5 µm) for 2–6 h before recording. After baseline recording, slices were perfused with control medium or ryanodine (20 µm), a selective ryanodine receptor blocker that is well established to inhibit CICR. Table 3 shows there were no effects of rapamycin on basic resting membrane parameters and measured potentials. These results are consistent with previous studies in hippocampal neurons in which no effect of rapamycin was found on voltage-gated Na+ and K+ currents or action potentials (Norris et al., 2002; Rüegg et al., 2007). However, baseline EPSP amplitude was reduced in rapamycin-treated cells (Table 3), an effect similar to that found in aged animals (Thibault et al., 2001; Burke and Barnes, 2006).
As seen in Figure 5, a significant enhancing effect of rapamycin was found on AHP measures. Rapamycin increased sAHP amplitude (p < 0.0001) and sAHP duration (p < 0.0001). Furthermore, a correlation analysis between incubation time and AHP measures revealed that this effect was time dependent because sAHP magnitude increased as a function of rapamycin incubation duration up to 6 h (sAHP amplitude vs incubation time, r = 0.74; p < 0.01). Treatment with ryanodine fully blocked the effect of rapamycin on AHP measures, indicating that the enhancing effect of rapamycin on AHP magnitude resulted primarily from elevated Ca2+ release from RyRs, as is also the case for the aging-related increase in the sAHP (Gant et al., 2006). Moreover, as reported previously (Kumar and Foster, 2005; Gant et al., 2006), ryanodine alone significantly reduced sAHP amplitude (p < 0.0001) and duration (p < 0.0001) compared with controls. A previous study reported a stimulatory effect of rapamycin on Ca2+-dependent K+ currents (Terashima et al., 1998). However, given the present data, that was likely an indirect effect resulting from FKBP1b disruption and increased CICR. Thus, our results indicate that rapamycin substantially increased sAHP magnitude by enhancing CICR in young-adult rat neurons, mimicking the effects of aging or Fkbp1b knockdown.
To more directly test the actions of rapamycin on CICR, we analyzed voltage-activated Ca2+ responses in a subset of the slice neurons in which the AHP was assessed. In previous studies of pyramidal neurons, it has been found that voltage-activated [Ca2+]i transients are increased with aging (Thibault et al., 2001; Hemond and Jaffe, 2005; Gant et al., 2006), apparently because of enhanced CICR (Gant et al., 2006). Ratiometric Ca2+ indicators were used in those previous studies to allow measures of absolute concentrations. Although single wavelength indicators do not generally provide information on absolute Ca2+ concentration values, they are useful for assessing Ca2+ responses relative to baseline (Jaffe et al., 1992; Miyakawa et al., 1992; Brown and Jaffe, 1994). Here, we used the single wavelength high-affinity fluorescent Ca2+ indicator Calcium Orange for improved signal/noise ratio and sampling speed (Fig. 6). To confirm comparable indicator loading, gray values were measured at baseline before obtaining stimulated fluorescence values. The baseline and post-ryanodine baseline gray values for nontreated control neurons were 206.9 ± 47.7 (mean ± SEM) and 209.3 ± 36.7 and for rapamycin-treated neurons were 232.4 ± 54.6 and 242.4 ± 58.7, indicating that there were no significant differences within or between treatment groups in indicator loading or baseline Ca2+. Postsynaptic intracellular Ca2+ transients were induced by 10 s trains of repetitive (7 Hz) synaptic stimulation at threshold for action potential generation (Fig. 6A1,A2), as described previously (Thibault et al., 2001; Gant et al., 2006). Rapamycin did not affect the RSS-activated peak fluorescence transient but increased the area under the curve (AUC) of the transient during the first 5 s of RSS (p = 0.05), primarily because of a near-significant decrease in rise time (p = 0.07) (Fig. 6B). The ryanodine-sensitive component (Fig. 6A1,A2) accounted for most of the enhancing effect of rapamycin on the AUC of the overall synaptically stimulated Ca2+ response, as shown by a significant increase specifically in the AUC of the ryanodine-sensitive component (p < 0.05) (Fig. 6C). Furthermore, the blocking effect of ryanodine indicates that the enhancement by rapamycin of Ca2+ transients, as of the AHP noted above, resulted primarily from enhancement by rapamycin of RyR-mediated CICR.
The results of these studies show that disruption of FKBP1b/12.6 function, using interfering RNA or rapamycin, induced key components of the aging phenotype of Ca2+ dysregulation in young hippocampal pyramidal neurons. Notably, both FKBP1b knockdown and rapamycin treatment increased VGCC activity in vitro and enhanced the sAHP in vivo. Thus, both approaches to disrupting FKBP1b function induced two of the hallmark biomarkers of aging-related Ca2+ dysregulation in young rat hippocampal pyramidal neurons. Moreover, additional studies showed that the enhancement by rapamycin of the sAHP and the Ca2+ transient was mediated primarily by increasing the ryanodine-sensitive (i.e., CICR) component of those responses, a mode of action highly analogous to that seen with aging (Gant et al., 2006).
This is the first study to implicate declining function of FKBPs as a candidate molecular mechanism in hippocampal Ca2+ dysregulation. It therefore supports the novel hypothesis that an aging-related decline in hippocampal FKBP1b/1a function plays a key role in unhealthy brain aging (Fig. 7). In addition, the present study provides the first direct evidence that selective FKBP1b disruption destabilizes Ca2+ responses in brain neurons.
Several other intriguing observations emerge from these studies. For one, the present data show that, in addition to their established effects on Ca2+ release from RyRs in cardiomyocytes (Zalk et al., 2007), FKBPs in neurons appear to regulate Ca2+ influx via VGCCs (Fig. 1). The VGCC lies upstream of the RyR in the VGCC–RyR circuit that amplifies Ca2+ responses through CICR (Berridge, 1997; Verkhratsky, 2004), and it is feasible that FKBPs regulate Ca2+ currents through their actions on RyRs and secondary retrograde effects on VGCCs. However, a more parsimonious explanation may be that FKBPs regulate VGCC activity directly.
Furthermore, RyRs are closely juxtaposed to L-VGCCS in cardiomyocytes (Sedarat et al., 2000; Protasi, 2002) and neurons (Chavis et al., 1996; Kim et al., 2007; Berrout and Isokawa, 2009), an organization that enables L-VGCCs to serve as a preferred source of Ca2+ for activating RyR-mediated CICR. Thus, considering the evidence that FKBP1b may directly modulate plasmalemmal VGCCs (Fig. 1) as well as RyRs, we suggest that, in neurons, FKBP molecules may function together with L-VGCCs and RyRs in a multimeric Ca2+ regulating complex in which the FKBP directly stabilizes both the RyR and L-VGCC via separate interactions (see model in Fig. 7).
Another striking observation in these studies is that the expression of FKBPs and of their inhibitory targets, RyR2 and mTOR (also called FRAP1, for FK506–rapamycin-associated protein) may be inversely related. That is, knockdown of FKBP1b appeared associated with upregulation of mTOR/FRAP1 and RyR2 (Fig. 4; Table 2), two inhibitory targets of FKBPs (see below). An inverse relationship could account for why we found downregulated expression of Fkbp1b and Fkbp1a in conjunction with upregulated Ryr2 and Frap1/Mtor in gene microarray studies of the hippocampus in aging rats (Kadish et al., 2009, their Table 3 and supplemental material). Additionally, there is some evidence of aging-related upregulation of the CaV1.3 poreforming subunit of the L-VGCC (Herman et al., 1998; Chen et al., 2000; Veng and Browning, 2002), which might also be an FKBP inhibitory target (Figs. 1, ,7).7). If additional research confirms this putative inverse relationship, it would suggest the interesting conclusion that FKBP1b and FKBP1a may inhibit several targets not only by physical interactions but by repression of gene/protein expression as well.
The relatively slow activation of Ca2+-related processes by FKBP knockdown or rapamycin (hours) is consistent with derepression of gene expression by FKBP1b disruption but could also reflect gradual dissociation of FKBPs from target molecules, among other possibilities. In contrast, FK506, another immunosuppressant drug, exerts inhibitory actions on VGCCs in minutes, apparently by inhibiting calcineurin (Norris et al., 2002).
FKBP1a and FKBP1b are members of the FKBP family of immunophilins, which bind the immunosuppressant drugs FK506 and rapamycin. Many immunophilins exhibit peptidyl-prolyl isomerase activity and function as chaperones and stabilizers (Marks, 1997; Eitoku et al., 2008; Kang et al., 2008; Jakob et al., 2009). FKBP1b and FKBP1a interact with multiple pathways and exert pleiotropic effects, including inhibition of the mTOR/FRAP1 pathway that plays an important role in cell growth and plasticity in the brain (Jacinto and Hall, 2003; Hoeffer et al., 2008). Thus, FKBPs bind and inhibit both RyRs and mTOR, albeit via separate pathways (Fig. 7A). Somewhat paradoxically, however, the FKBP ligand rapamycin has opposite effects on these two inhibitory pathways. Although binding of rapamycin weakens FKBP-mediated inhibition of Ca2+ release pathways in myocytes (Marks, 1997), rapamycin binding strengthens FKBP-associated inhibition of mTOR. Recently, such rapamycin-enhanced mTOR inhibition has been reported to extend lifespan in mice and counteract some aging markers and AD-model brain pathology, possibly because of suppression of aberrant growth or of insulin signaling and tumorigenesis (Harrison et al., 2009; Caccamo et al., 2010). Nevertheless, in some studies, rapamycin has impaired LTP and/or cognitive function (Tang et al., 2002; Cracco et al., 2005; Parsons et al., 2006; Blundell et al., 2008), and the results of the present study also raise a clear caveat regarding the potential use of rapamycin as an “anti-aging” therapy. In any event, the increased inhibitory actions of rapamycin on mTOR appear essentially unrelated to its opposite disinhibitory effects on Ca2+ regulation, as the latter mimic the effects of aging. Therefore, with regard to Ca2+ regulation, knockdown of Fkbp1b and rapamycin appear to provide complementary approaches for selective disruption of FKBP1b function.
The finding that disrupting hippocampal FKBP function can induce the Ca2+ dysregulation phenotype raises the question of whether similar disruption may occur with aging and, if so, what factors might underlie an initial aging-related decline in FKBP function. One possibility is that a decrease in the biosynthesis of energy-expensive proteins, notably including FKBPs, might result from a metabolic shift that appears to develop in neurons and glia of the hippocampus during aging (Rowe et al., 2007; Brinton, 2008; Kadish et al., 2009; Blalock et al., 2010). In addition, adrenal stress hormones (glucocorticoids), which have long been linked to brain aging (Porter and Landfield, 1998) and may exhibit altered impact with aging (Holmes and Seckl, 2006; Landfield et al., 2007; Blalock et al., 2010), enhance the Ca2+-dependent sAHP in the hippocampus (Joёls and de Kloet, 1989; Kerr et al., 1989). Similarly, mineralocorticoids have been shown to decrease FKBP1b expression and enhance RyR-mediated Ca2+ release (Gómez et al., 2009). Another possibility is that gradual vascular insufficiency might result in focally decreased FKBP protein (Kato et al., 2000). Alternatively, decreased FKBP function might be mediated by posttranslational events as well as by altered expression. In myocytes, hyperphosphorylation of RyR2 by protein kinase A (PKA) results in dissociation from FKBP1b and enhanced Ca2+ release by RyR2 (Marx et al., 2000). With aging, gene expression for PKA-pathway molecules appears increased in the hippocampus (Rowe et al., 2007; Kadish et al., 2009). However, there appear to be other pathways operating in the hippocampus because PKA activation does not increase but rather reduces the sAHP in hippocampal neurons (Oh et al., 2009). Clearly, substantial additional work will be needed to resolve these questions.
Viral-vector-mediated manipulation of gene expression in the adult CNS is a relatively new and powerful experimental tool that allows selective targeting of specific molecular pathways and cellular populations while circumventing some of the pitfalls of conventional genetic and pharmacological interventions (Fitzsimons et al., 2002; Klein et al., 2002, 2008; Zaiss and Muruve, 2005; Cearley and Wolfe, 2006; Mouravlev et al., 2006). In particular, it appears to provide an important complementary approach for gene array techniques because it can facilitate rapid and direct testing of hypotheses generated by genome-wide analyses of expression. AAV vector-mediated genomic intervention also appears to have substantial clinical potential, based on its persisting effects and its as yet relatively low pathogenic profile. In fact, several preclinical and clinical studies using virally mediated gene delivery for the treatment of neurodegenerative disorders have been conducted (Crystal et al., 2004; McPhee et al., 2006; Kaplitt et al., 2007; Johnston et al., 2009; Su et al., 2009; Kells et al., 2010). Thus, in concert with previous findings showing that FKBP1b expression in the hippocampus of incipient AD subjects is decreased and correlates with impaired cognition (Blalock et al., 2004), the present results on virally mediated manipulation of FKBP function may have future therapeutic implications.
In summary, we found that disrupting hippocampal FKBP1b function or expression was sufficient to induce aging-like Ca2+ dysregulation. Together, the results support a novel hypothesis of unhealthy brain aging, shown in Figure 7. In this model, an aging-related decline in hippocampal FKBP1b function, which can be mimicked by FKBP knockdown or rapamycin, leads to destabilization of both L-VGCCs and RyRs, resulting in elevated Ca2+ influx, greater CICR, larger sAHPs, and impaired excitability/plasticity and memory. If additional studies support this model, FKBP-mediated Ca2+-regulatory pathways may become important new therapeutic targets for combating aging-related cognitive impairment and neurodegenerative disease.
This work was supported by National Institute on Aging Grants AG004542 and AG010836 and National Center for Research Resources Grant P20-RR15592.